Before You Begin
Before you restore workloads to Google Compute Engine, consider the following requirements and limitations:
- Check whether a helper appliance must be configured for restore. For more information, see Helper Appliances.
- The backup server and repositories with workload backup files must have access to the internet.
If backup files are located on deduplicating storage appliances or shared folder repositories, the internet connection is required for gateway servers that communicate with these repositories.
- If you use a cloud-init-based Linux distribution, we recommend that you use SSH keys on these distributions. If you use a password, it is blocked after restore for security reasons. To reset the password on the restored VM, use the technologies described in Google Cloud Documentation.
- You must have a backup of the workload that you plan to restore to Google Compute Engine.
- Make sure that the Cloud Build API is enabled. For more information on enabling the Cloud Build API and other requirements for importing virtual disks into Google Compute Engine, see the Google Cloud Documentation.
- Make sure the IAM service account that you plan to use to restore workloads to Google Compute Engine has permissions to restore to Google Compute Engine. For more information, see Google Compute Engine IAM User Permissions.
- If you restore workloads from backups of virtual and physical machines (non-Google Compute Engine virtual machines), check the supported operating systems and their differences from standard images in the Google Cloud documentation.
- Check that the logical sector size of disks that you plan to restore is less than 4096 bytes. Contents of disks whose logical sector size is 4096 bytes will be unreadable in Google Compute Engine.
- If you plan to assign Google labels to the restored workload, check limitations for labels in the Google Cloud documentation.